Rebound effects

purchase green products, and did not adopt them in the long term when they did.

Peyer at al. (2017) studied voluntary simplicity and found that almost sixth of the German population are voluntary simplifiers; they buy more green products and have greater consciousness in terms of environmental and economic sustainability than the other four segments found. They identified the voluntary simplifiers based on their households’

consumption levels, which were measured by the number of 11 consumer goods, such as cars, smart phones, and skis, in their homes, and their monthly net household income adjusted according to the number of adults and children in the household. Groups that were neither voluntary simplifiers nor over-consumers showed a strong positive correlation between owned consumer goods and income. Voluntary simplifiers had a relatively low number of consumer goods related to their income, and over-consumers had a high number. The segments of less well-off consumers couldn’t afford different consumption choices as they focused primarily on bare necessities.

Most EOA studies have been published in business or psychology journals, and studies published in environmental journals have mostly focused on motivations, attitudes, reasons, and anti-consumption behaviour. (Garcia-de-Frutos et al., 2018). Touchette and Nepomuceno (2020) examined the environmental impact of anti-consumption lifestyles (voluntary simplicity, frugality, and tightwadism), environmental concern, and ethically minded consumption. They calculated respondents’ carbon footprints based on information collected from them and presented a questionnaire to assess anti-consumption lifestyles and environmental and ecological concerns. The results were similar to those of Kropfeld et al. (2018), indicating that tightwadism can be associated with lower GHG emissions. Tightwads with higher knowledge of emission effects have lower GHG emissions; their desire to avoid spending causes them to consume significantly less. The results indicated that there was no correlation between environmental concerns/voluntary simplicity/frugality and positive impact on environment. Rich et al. (2020) did not study GHG emissions, but their findings are similar to those of Touchette and Nepomuceno (2020) in that they found no difference between voluntary simplifiers and non-simplifiers in terms of finding environmental important.

2.4

The rebound effect is a phenomenon that occurs when achieved gains are partly or completely offset by increased use, such as when improvements in energy efficiency lead to an increased use of electricity. Rebound effects can be separated into direct and indirect effects in the context of microeconomies like households. Direct effects, in terms of energy efficiency, are created when cheaper energy increases the overall demand for energy. Indirect effects occur when cheaper energy increases the demand for other goods and services, which leads to increased GHG emissions in other sectors. (Chitnis, 2013;

Druckman et al., 2011). Direct and indirect rebound effects can both be further divided

2 Theoretical background 28

into income and substitution effects. Income effects occur when improved energy efficiency increases the real income of households through cheaper energy bills, which leads to increased consumption overall. Substitution effects, on the other hand, occur when households’ real income remains constant, and they shift their consumption of a particular service or good to a similar but differently priced service or good (Chitnis et al., 2013; Investopedia, 2020). The breakdown is theoretical in the context of GHG emissions, and the result is the sum of these two effects. In addition to these micro effects, there are macro effects that result from the interaction between consumers and producers.

Secondary effects occur when, for example, an energy efficiency measure reduces costs for an industry, leading to a decrease in the prices of goods or services. This, in turn, leads to an increased demand for these goods and services, and thus also in energy. Economy-wide effects occur when the demand for fuel decreases due to increased energy efficiency;

the price reduction then leads to increasing amounts of purchased fuel. Transformational effects happen when technology changes have the potential to, “change consumer preferences, alter social institutions, and rearrange the organization of production,”

(Hertwich, 2005). It is relevant to acknowledge the macroeffects, though this dissertation focuses on households, i.e. microeffects.

Chitnis et al. (2013) modelled how cost savings from seven energy efficiency measures in UK dwellings would be spent across different consumption categories; they included both direct and indirect effects. The range of rebound effect was 5–15%, and the main source of rebound effects was spending cost savings on non-energy related goods and services. The rebound effect stayed moderate, as these services were less GHG intensive than energy production. Similarly, Druckman et al. (2011) estimated the rebound effect to be 7% when lowering the room temperature by 1 ℃ in the UK context. The results of both studies were highly depended on the GHG emissions of UK energy production. In countries with lower energy-related GHG emissions, the rebound effect would be greater.

Additionally, substitution effects might present a greater rebound effect depending on what the cost savings would be spent on (Chitnis et al., 2013).

Druckman et al. (2011) found the rebound effects to be significantly larger when eliminating food waste, thus reducing food expenditure by a third (51%) and for walking or cycling instead of driving a car for a trips of two miles or less (25%). In the study, savings deposited into a bank account were treated as investments and an average GHG intensity for UK investments was used. In a behaviour-as-usual scenario, 4% of the savings were invested and the rest were re-spent. If the 7% rebound effect from lowering room temperature is included, the total rebound effect for these three actions becomes 34%. They also estimated the “least-worst” rebound effect, i.e. savings used in the category of housing (household rent, maintenance, repair, and water supply), which had the lowest GHG intensity of all the consumption categories. In this case, the rebound effect was 12%. Accordingly, they also estimated the worst-case rebound effect, in which the savings were used for gas. This resulted in an extreme backfire; rebound rate of 515%.

They also investigated how the savings ratio would influence the rebound effect. The lowest savings rate in the UK between 1964 and 2009 was –4%, meaning that households were withdrawing from savings; the rebound effect in this case was 35%. With a high

2.4 Rebound effects 29 savings ratio of 40%, the rebound effect was 31%. Assuming that all savings were invested, the rebound effect would be 26%. The difference comes from investments having a slightly lower GHG intensity than consumption expenditures.

Similarly, Chitnis et al. (2014) assumed that households saved and invested 15% of their annual income and used an UK average for the GHG intensity. According to their calculations, indirect rebound effects account for majority of GHG emissions. Embodied emissions of non-energy goods and services had the greatest impact, though a larger share of rebound effects could be attributed to direct emissions in cases of low-income households. They found that rebound effects were generally larger for low-income households due to these households spending cost savings on GHG-intensive goods like food. Murray (2013) found similar results regarding lower income households, though he still suggested targeting changes in consumer behaviour, especially conservation measures, toward higher income households.

Ottelin et al. (2017) focused on the Finnish working middle class and studied the rebound effects of reduced driving and car ownership and compared car owners to car-free households, keeping the characteristics otherwise similar. They found the rebound effect for giving up car ownership to be 68%, whereas the average rebound effect for reduced driving was 23%. Persons who own a car but drive very little were found to have the lowest carbon footprint in terms of transportation; it was estimated to be 11% lower than similar persons who do not own a car. This implies that money saved by not owning a car is directed into other consumption categories.

Font Vivanco et al. (2014) developed a general microeconomic model to study the environmental rebound effect of plug-in hybrid cars, full-battery electric cars, and hydrogen fuel cell cars. They combined LCA-based methods with a marginal consumption model based on technology choices. In terms of GHG emissions, they found a rebound effect of less than 5% for a plug-in hybrid, and a notable negative rebound effect for a full-battery electric and hydrogen fuel cell cars. The moderate rebound effect of a hybrid was due to a slight decrease in transport costs as compared to its alternative.

The negative rebound effect for a full-battery car was caused by high capital costs, leaving less income for other consumption categories. The GHG emissions of using a full battery electric car were found to be 79% smaller as compared to spending the same amount of money on general consumption. In both cases, green production technologies were also named as a factor in the reduced GWP impacts. The results were also analyzed across different income quintiles; lower income groups were found to have a higher rebound effect, as freed income is generally spent on categories with higher environmental impacts.

Similar results were achieved by Mizobuchi (2008); they showed a significantly lower rebound effect when the capital costs were considered. Without considering capital costs, the rebound effect was 115%. When capital costs were considered, it was 27%. The study took electric appliances, such as air conditioners, TVs, burners, heaters, and cars, into consideration. The study indicated that most energy-efficient appliances were more

2 Theoretical background 30

expensive than less efficient ones. Chitnis et al. (2013) presented similar results; solar thermal heating and LED lightning were found to have a negative rebound effect when capital costs were considered. Similarly, Ottelin et al. (2015) suggested that the smaller carbon footprints of households living in new housing (as compared to older housing in similar area) are due to higher housing loans, leaving not as much money for other consumption. However, their results show that the carbon footprints of households living in new housing are higher as compared to older housing in inner urban areas. In these cases, high levels of other consumption counteracted the energy efficiency gains.

31

3 Materials and methods

This chapter presents the materials and methods used in this thesis. First, the research approach is discussed, and overview of the methods, data collection techniques, and analysis used follow. More detailed descriptions are provided in individual articles.

Finally, the double impact framework created in Publication II is discussed.

3.1

Quantitative research conventionally produces numbers and percentages which can be presented as “facts” at least within the given sample, whereas qualitative research is used in answering questions with deeper insight (Barnham, 2015). Due to the research questions of the thesis which mainly require numerical answers, a quantitative approach was selected. A multimethod approach was seen to be the most suitable method for this dissertation as various quantitative analysis were needed. Mixed methods are sometimes seen as synonymous with multimethods, and sometimes a clear distinction is created, generating confusion (Anguera et al., 2018). The prevailing consensus, however, is that, in multimethod approach, complementary methodologies are used to answer the research goal; there is not necessarily a difference in terms of whether the methodologies are quantitative, qualitative, or both. Conversely, both quantitative and qualitative methods are applied in mixed methods studies (Hunter & Brewer, 2015; Anguera et al., 2018). In this dissertation, quantitative methods were used in forms of calculations based on statistical data, questionnaires, and life cycle assessments.

The context for all publications was Finland, while the focus varied across publications (Table 1). The main research goal of the dissertation is divided into sub-questions.

Publications I–III contribute to more than one sub-question, while Publication IV contributes to only one. Publications I and II focus on overall household consumption, and Publication II is partially built on the results from Publication I. Publication III focuses more specifically on low-carbon housing and Publication IV focuses on knowledge of and willingness to take climate change mitigation actions.

3 Materials and methods